Figure 1. Average summer location of the Bering Sea cold pool (temperature < 2° C) during the first 5 years (1982-86) and most recent 5 years (2002-06) of the RACE Division eastern Bering Sea bottom trawl survey time series.
Sea ice is a dominant factor in the year-round bottom water conditions on the Bering Sea shelf. This layer of winter ice creates a pool of cold (< 2° C) bottom water that persists through the summer. Loss of sea ice from the Bering Sea therefore implies a loss of the cold water pool with a greater warming of bottom temperatures and subsequent potential for concomitant ecological reorganization of marine organisms.
Program scientist Mike Litzow and collaborator Franz Mueter (Sigma-Plus Consulting), with support from the North Pacific Climate Regimes and Ecosystem Productivity (NPCREP) study, analyzed the RACE Division Bering Sea bottom trawl survey time series (1982-2006) to test for northward shifts in the location of the cold pool and the centers of distribution of both fish and crustacean populations in the southeast Bering Sea.
Their analysis found that the average bottom temperatures measured during the survey period, when adjusted for seasonal differences in sampling, warmed approximately 0.9°C since 1982 (from 2.1° to 3.0°C) as defined by the best linear fit to annual mean values. At the same time, the southern edge of the cold pool (defined by the 2°C isotherm) has shifted approximately 230 km northwards, from an average southernmost point of about lat. 56.0°N in 1982-86 to lat. 58.1°N in 2002-06 (Fig. 1).
As would be expected, shifts in distribution of marine organisms would be a conspicuous biological response to these changes in ocean conditions, and their analysis showed the average center of abundance for 45 of the most common fish and crustacean taxa sampled on the trawl survey has migrated 31 ± 60 (SD) km northwards since 1982. Considerable variability exists among taxa in the degree of distribution change, and we could not explain this variability in terms of commercial fishing effects, trophic level, foraging habitat, or life history traits.
Figure 2. Warming climate and community-wide distribution shifts in the Bering Sea, 1982-2006. (A) Direct temperature effects on community-wide centers of abundance. Each dot represents peak center of abundance averaged across 45 fish and crustacean taxa for one survey year. Distance values on y-axis reflect difference from mean latitude of community-wide centers of abundance averaged across entire time series. (B) Trend in residual community-wide centers of abundance after direct temperature effects from (A) have been removed.
Figure 3. Immediate management implications of receding Bering Sea sea ice. (A) Correlation between ice extent and commercial snow crab catch, 1982-2005. (B) Correlation between ice extent and mean trophic level of entire Bering Sea commercial catch, 1982-2004. Ice cover index in both panels has been smoothed with 3-year running mean. Solid line indicates smoothed non-parametric regression, dotted lines indicate 95% confidence intervals around best-fit regression.
Direct temperature effects explain a large degree of observed community-wide northward shifts in abundance (Fig. 2A), but residual variability in centers of distribution not explained by bottom temperature shows a coherent temporal trend towards accelerating distribution change in recent years (Fig. 2B).
This residual variability was not explained by the other climate parameters that we examined (e.g., ice cover, sea level pressure, wind mixing, alongshore wind stress). One of the great challenges in forecasting ecological responses to climate forcing is the potential for emergent ecological effects that magnify the effect of climate perturbation. Franz and Mike speculate that the trend in residual variability in community-wide distribution change (Fig. 2B) might indicate an emerging effect whereby the sum of community-wide distribution change exceeds that expected from direct climate effects.
Understanding the causes of this trend in residual variability and variability among taxa in distributional responses to warming are two important research challenges for understanding the response of the Bering Sea ecosystem to future warming.
Finally, the loss of sea ice and warming of bottom waters has immediate management implications. Commercial catches of snow crab (Chionoecetes opilio), far and away the most important commercial species in the Bering Sea arctic community, are strongly positively correlated with the extent of sea ice (Fig. 3A).
Recession of sea ice in recent decades also has implications for the sum total of commercial Bering Sea fisheries. The average trophic level of all Bering Sea fisheries is negatively correlated with ice extent (Fig. 3B), indicating a reorganization in fisheries as ice
has retreated, where crab catches have declined and groundfish catches have stayed constant or increased.